#!/usr/bin/perl
############################################################
#
# File: addates.pl
#
# Arguments: filespec -s
#   filespec  file specification to process (if none, then assumes all)
#   -s        if specified, recurse subdirectories
#
# Description:
# Adds a date field to W3C log files that don't have them
# Assumes that all log entries are for the date specified
# in the header of the file.
#
# NOTE: This only works for log files rotated at midnight each day!
#       No longer true (see mod 2000Mar01) but assumes that there are
#       no 24hour gaps in the logfile
#
# WARNING: This changes the original files! Backup first!
#
# Created: 12.February.1999 by Jeremy Wadsack (jwadsack@wadsack-allen.com)
# 	Copyright (C) 1999 Wadsack-Allen. All rights reserved.
# Modified: 1.March.2000 by Guido Van Hoecke (Guido.VanHoecke@advalvas.be)
#       No copyrights: released in the public domain
#
# This script is free software; you can redistribute it and/or
# modify it under the same terms as Perl itself.
#
##########################################################################
# Date      Modification                                            Author
# ------------------------------------------------------------------------
# 1999Apr09 Included Expand funciton instead of module.                 JW
# 2000Mar01 Added NextDate function and logic.                         GVH
# 2000May31 Allows for multiple #Field lines (but note that it ignores
#           multiple #Date: lines because these are likely wrong).      JW
# 2000May31 Optimized by removing extra loop on contents.               JW
##########################################################################

use Time::Local;

# If no arguments, then assume all files
if( @ARGV == 0 ) {
    $ARGV[0] = '*.*';
} # end if
# If -s then recurse subdirectories
if( $ARGV[0] =~ /\-s/i ) {
    $ARGV[1] = $ARGV[0];
    $ARGV[0] = '*.*';
} # end if

my @filelist = Expand_Files( $ARGV[0], $ARGV[1] =~ /\-s/i );

foreach $filename (@filelist ) {
    # Open file and get contents (!WARNING this uses lot of memory)
    print "Checking file $filename. ";
    open( INPUT, $filename );
    @contents = <INPUT>;
    close( INPUT );
    unless( @contents ) {
	print "File is empty. Skipping.\n";
	next;
    } # end unless

    # -- Find the date in the first pass, so we can
    #    abort if there is none
    $Date = '_empty_';
    foreach $contents (@contents) {
	if( $contents =~ /^\#Date: (.{10})/i ) {
	    $Date = $1;
	    print "Date $Date. ";
	    last;
	} # end if
	# -- Jump out of loop if not a starting spec line (just a little OPT)
	last if ( $contents =~ /^[^\#]/ );
    } # end foreach
    if( $Date eq '_empty_' ) {
	print "File has no date. Skipping.\n";
	next;
    } # end if

    # -- Find and update Field specification and replace all log file lines
    $skipFile = 0;
    print "Processing... ";
    $previousTime = '00:00:00'; #smallest possible time
    foreach $contents (@contents) {
	if( $contents =~ /^\#Fields:(.+)$/i ) {
	    $Spec = $1;
	    if( $Spec =~ /date/i ) {
		# -- File already has a date field so skip the file
		$skipFile = 1;
		print "File has date field. {$Spec} Skipped\n";
		last
	    } else {
		$contents = "#Fields: date$Spec\n";
	    } # end if
	} elsif( $contents =~ /^[^\#]/ ) {
	    $contents =~ /(\S+)\s+/; # match first word, i.e. the time
	    $thisTime = $1;
	    if ( $thisTime lt $previousTime ) {
		# date has changed
		$Date = NextDate( $Date );
		print "Date $Date. ";
	    } #end if
	    $previousTime = $thisTime;
	    $contents = "$Date $contents";
	} # end if
    } # end foreach
    next if $skipFile;
    
    # Write new file to same name
    print "done.\n";
    open( OUTPUT, ">$filename" );
    print OUTPUT @contents;
    close( OUTPUT );

} # end foreach



#------------------------------------
# Sub: Expand_Files
#
# Arguments: $filespec, $cascade
#
# Returns: @filelist
#
# Description:
#  returns a list of filenames based on the file
#  specification and the choice to cascade/resurse
#  subdirectories.
#-------------------------------------
sub Expand_Files {
    my ( $spec, $cascade ) = @_;
    my (
	@filenames,
	$dir, $filespec, $file,
	@dirnames, $dirs
	);

    @filenames = ("");

    if(( $spec =~ /.*[\*\?].*/ ) || $cascade) {
	# expand the set
	if( !( ($dir, $filespec) = ( $spec =~ /^(.*)[\\\/](.+)$/ ) ) ) {
	    $filespec = $spec;
	    $dir = '.';
	} # end if
	# mung any re metacharacters to literals
	quotemeta( $filespec );
	# mung the wildcards to re's
	$filespec =~ s/\*/\.\*/;
	$filespec =~ s/\?/\./;
	# cascade if enabled
	if( $cascade != 0 ) {
	    #recurse the directory list
	    opendir(DIR, $dir) || die "Error: Can't read directory $dir: $!";
	    @dirnames = grep { -d "$dir/$_" } readdir(DIR);
	    closedir DIR;
	    foreach $dirs( @dirnames ) {
		@filenames = (@filenames, &Expand_Files( "$dirs/$filespec" ));
	    } # end foreach
	} # end if
	opendir(DIR, $dir) || die "Error: Can't read directory $dir: $!";
	@filenames = (@filenames, grep { /$filespec/ && -f "$dir/$_" } readdir(DIR));
	closedir DIR;
	foreach $file( @filenames ) {
	    $file = "$dir/$file";
	} # end foreach
    } else {
	# Just a single file, return it
	@filenames = ( $spec )
	} # end if
    return( @filenames );
} # end Expand_Files


#------------------------------
# Sub: NextDate
#
# Arguments: $startDate (as yyyy-mm-dd)
#
# Returns: $nextDate
#
# Description: increments the $startDate variable to obtain $nextDate
#
#------------------------------
sub NextDate {
    my ( $startDate ) = @_;
    my ( $year, $mon, $mday, $time, $nextDate );
    $startDate =~ /(\w+)\W+(\w+)\W+(\w+)/;
    $year = $1 - 1900;
    $mon = $2 - 1;
    $mday = $3;
    $time = timelocal(0, 0, 12, $mday, $mon, $year);         # Noon
    $time += 86400;                                          # Tomorrow
    ($mday, $mon, $year) = (localtime $time)[3..5];
    $nextDate = sprintf "%04d-%02d-%02d", 1900 + $year, $mon + 1, $mday;
  }